Macromolecular modeling and design in Rosetta: recent methods and frameworks.

The Rosetta software for macromolecular modeling, docking and design is extensively used in laboratories worldwide. During two decades of development by a community of laboratories at more than 60 institutions, Rosetta has been continuously refactored and extended. Its advantages are its performance and interoperability between broad modeling capabilities. Here we review tools developed in the last 5 years, including over 80 methods. We discuss improvements to the score function, user interfaces and usability. Rosetta is available at http://www.rosettacommons.org.

Modeling membrane geometries implicitly in Rosetta.

Interactions between membrane proteins (MPs) and lipid bilayers are critical for many cellular functions. In the Rosetta molecular modeling suite, the implicit membrane energy function is based on a "slab" model, which represent the membrane as a flat bilayer. However, in nature membranes often have a curvature that is important for function and/or stability. Even more prevalent, in structural biology research MPs are reconstituted in model membrane systems such as micelles, bicelles, nanodiscs, or liposomes. Thus, we have modified the existing membrane energy potentials within the RosettaMP framework to allow users to model MPs in different membrane geometries. We show that these modifications can be utilized in core applications within Rosetta such as structure refinement, protein-protein docking, and protein design. For MP structures found in curved membranes, refining these structures in curved, implicit membranes produces higher quality models with structures closer to experimentally determined structures. For MP systems embedded in multiple membranes, representing both membranes results in more favorable scores compared to only representing one of the membranes. Modeling MPs in geometries mimicking the membrane model system used in structure determination can improve model quality and model discrimination.

Simultaneous prediction of protein secondary structure and transmembrane spans.

Prediction of transmembrane spans and secondary structure from the protein sequence is generally the first step in the structural characterization of (membrane) proteins. Preference of a stretch of amino acids in a protein to form secondary structure and being placed in the membrane are correlated. Nevertheless, current methods predict either secondary structure or individual transmembrane states. We introduce a method that simultaneously predicts the secondary structure and transmembrane spans from the protein sequence. This approach not only eliminates the necessity to create a consensus prediction from possibly contradicting outputs of several predictors but bears the potential to predict conformational switches, i.e., sequence regions that have a high probability to change for example from a coil conformation in solution to an α-helical transmembrane state. An artificial neural network was trained on databases of 177 membrane proteins and 6048 soluble proteins. The output is a 3 × 3 dimensional probability matrix for each residue in the sequence that combines three secondary structure types (helix, strand, coil) and three environment types (membrane core, interface, solution). The prediction accuracies are 70.3% for nine possible states, 73.2% for three-state secondary structure prediction, and 94.8% for three-state transmembrane span prediction. These accuracies are comparable to state-of-the-art predictors of secondary structure (e.g., Psipred) or transmembrane placement (e.g., OCTOPUS). The method is available as web server and for download at www.meilerlab.org.

Protein remote homology detection and structural alignment using deep learning.

Exploiting sequence-structure-function relationships in biotechnology requires improved methods for aligning proteins that have low sequence similarity to previously annotated proteins. We develop two deep learning methods to address this gap, TM-Vec and DeepBLAST. TM-Vec allows searching for structure-structure similarities in large sequence databases. It is trained to accurately predict TM-scores as a metric of structural similarity directly from sequence pairs without the need for intermediate computation or solution of structures. Once structurally similar proteins have been identified, DeepBLAST can structurally align proteins using only sequence information by identifying structurally homologous regions between proteins. It outperforms traditional sequence alignment methods and performs similarly to structure-based alignment methods. We show the merits of TM-Vec and DeepBLAST on a variety of datasets, including better identification of remotely homologous proteins compared with state-of-the-art sequence alignment and structure prediction methods.

Structure-based protein function prediction using graph convolutional networks.

The rapid increase in the number of proteins in sequence databases and the diversity of their functions challenge computational approaches for automated function prediction. Here, we introduce DeepFRI, a Graph Convolutional Network for predicting protein functions by leveraging sequence features extracted from a protein language model and protein structures. It outperforms current leading methods and sequence-based Convolutional Neural Networks and scales to the size of current sequence repositories. Augmenting the training set of experimental structures with homology models allows us to significantly expand the number of predictable functions. DeepFRI has significant de-noising capability, with only a minor drop in performance when experimental structures are replaced by protein models. Class activation mapping allows function predictions at an unprecedented resolution, allowing site-specific annotations at the residue-level in an automated manner. We show the utility and high performance of our method by annotating structures from the PDB and SWISS-MODEL, making several new confident function predictions. DeepFRI is available as a webserver at https://beta.deepfri.flatironinstitute.org/ .

Integrative Protein Modeling in RosettaNMR from Sparse Paramagnetic Restraints.

Computational methods to predict protein structure from nuclear magnetic resonance (NMR) restraints that only require assignment of backbone signals, hold great potential to study larger proteins. Ideally, computational methods designed to work with sparse data need to add atomic detail that is missing in the experimental restraints. We introduce a comprehensive framework into the Rosetta suite that uses NMR restraints derived from paramagnetic labeling. Specifically, RosettaNMR incorporates pseudocontact shifts, residual dipolar couplings, and paramagnetic relaxation enhancements. It continues to use backbone chemical shifts and nuclear Overhauser effect distance restraints. We assess RosettaNMR for protein structure prediction by folding 28 monomeric proteins and 8 homo-oligomeric proteins. Furthermore, the general applicability of RosettaNMR is demonstrated on two protein-protein and three protein-ligand docking examples. Paramagnetic restraints generated more accurate models for 85% of the benchmark proteins and, when combined with chemical shifts, sampled high-accuracy models (≤2Å) in 50% of the cases.

Mapping genetic variations to three-dimensional protein structures to enhance variant interpretation: a proposed framework.

The translation of personal genomics to precision medicine depends on the accurate interpretation of the multitude of genetic variants observed for each individual. However, even when genetic variants are predicted to modify a protein, their functional implications may be unclear. Many diseases are caused by genetic variants affecting important protein features, such as enzyme active sites or interaction interfaces. The scientific community has catalogued millions of genetic variants in genomic databases and thousands of protein structures in the Protein Data Bank. Mapping mutations onto three-dimensional (3D) structures enables atomic-level analyses of protein positions that may be important for the stability or formation of interactions; these may explain the effect of mutations and in some cases even open a path for targeted drug development. To accelerate progress in the integration of these data types, we held a two-day Gene Variation to 3D (GVto3D) workshop to report on the latest advances and to discuss unmet needs. The overarching goal of the workshop was to address the question: what can be done together as a community to advance the integration of genetic variants and 3D protein structures that could not be done by a single investigator or laboratory? Here we describe the workshop outcomes, review the state of the field, and propose the development of a framework with which to promote progress in this arena. The framework will include a set of standard formats, common ontologies, a common application programming interface to enable interoperation of the resources, and a Tool Registry to make it easy to find and apply the tools to specific analysis problems. Interoperability will enable integration of diverse data sources and tools and collaborative development of variant effect prediction methods.

Capturing conformational States in proteins using sparse paramagnetic NMR data.

Capturing conformational changes in proteins or protein-protein complexes is a challenge for both experimentalists and computational biologists. Solution nuclear magnetic resonance (NMR) is unique in that it permits structural studies of proteins under greatly varying conditions, and thus allows us to monitor induced structural changes. Paramagnetic effects are increasingly used to study protein structures as they give ready access to rich structural information of orientation and long-range distance restraints from the NMR signals of backbone amides, and reliable methods have become available to tag proteins with paramagnetic metal ions site-specifically and at multiple sites. In this study, we show how sparse pseudocontact shift (PCS) data can be used to computationally model conformational states in a protein system, by first identifying core structural elements that are not affected by the environmental change, and then computationally completing the remaining structure based on experimental restraints from PCS. The approach is demonstrated on a 27 kDa two-domain NS2B-NS3 protease system of the dengue virus serotype 2, for which distinct closed and open conformational states have been observed in crystal structures. By changing the input PCS data, the observed conformational states in the dengue virus protease are reproduced without modifying the computational procedure. This data driven Rosetta protocol enables identification of conformational states of a protein system, which are otherwise difficult to obtain either experimentally or computationally.